Reproducible visual-enhancer example with MIT CLI and real pair
#1
by shi9214 - opened
I put together a different, minimal example for runtime and ComfyUI researchers: https://huggingface.co/datasets/shi9214/dlss5-visual-enhancer-workflow. It is deliberately small but includes actual files, a transparent input/result pair, and a MIT-licensed command-line dry run.
Source: 
Output: 
python3 enhance_workflow.py --input examples/input.jpg --dry-run --output request.json
The project is independent and not affiliated with NVIDIA; it does not present itself as an official DLSS implementation and publishes no unmeasured FPS or quality claims. See https://www.dlss5nvidia.com/visual-enhancer for the user-facing visual-enhancer context and known limitations. If this format is useful, the dataset card shows how to extend it with your own licensed examples.